Welcome to DU! The truly grassroots left-of-center political community where regular people, not algorithms, drive the discussions and set the standards. Join the community: Create a free account Support DU (and get rid of ads!): Become a Star Member Latest Breaking News Editorials & Other Articles General Discussion The DU Lounge All Forums Issue Forums Culture Forums Alliance Forums Region Forums Support Forums Help & Search

Judi Lynn

(163,392 posts)
Mon Apr 28, 2025, 01:29 AM 14 hrs ago

Google Is Training a New A.I. Model to Decode Dolphin Chatter--and Potentially Talk Back

Last edited Mon Apr 28, 2025, 03:21 AM - Edit history (1)

The company says its new model, called DolphinGemma, will be made open source this summer. Researchers are also trying to train dolphins to mimic made-up names for certain objects

Sara Hashemi - Daily Correspondent

April 25, 2025

The company says its new model, called DolphinGemma, will be made open source this summer. Researchers are also trying to train dolphins to mimic made-up names for certain objects



A newly developed A.I. model is based on 40 years of vocalizations from a community of Atlantic spotted dolphins. Google

Dolphins are clever communicators. The animals use complex clicks, squawks and whistles to call out to each other, fight and attract a mate. Now, Google says it is developing a large language model (LLM) that can make better sense of those vocalizations—and, maybe, allow humans to talk back.

Over the last 40 years, researchers at the Wild Dolphin Project (WDP) have collected audio and video of a community of Atlantic spotted dolphins (Stenella frontalis) in the Bahamas. The new artificial intelligence model, called DolphinGemma, is trained on that database. It listens to the dolphins’ vocalizations, identifies patterns and predicts what comes next, just as LLMs do with human language.

Using A.I. “could give us the opportunity to see patterns that, from a human perspective, we may not look at,” says Thea Taylor, who manages the Sussex Dolphin Project in England and is not involved with Google’s work, to Melissa Hobson at Scientific American. Relying on the LLM can also speed up the data analysis process, which would take a human more than 100 years by hand.



Researchers with Google, the WDP and Georgia Tech are also working on a device called CHAT, or Cetacean Hearing Augmentation Telemetry. The wearable technology will allow researchers to generate dolphin-like sounds made up by A.I. to refer to specific items that dolphins enjoy, like seagrass or sargassum.

More:
https://www.smithsonianmag.com/smart-news/google-is-training-a-new-ai-model-to-decode-dolphin-chatter-and-potentially-talk-back-180986434/
Latest Discussions»Culture Forums»Science»Google Is Training a New ...